Cloning and localization of the in vitro functional origin of replication of bacteriophage T7 DNA.

Segments of a Hpa I fragment which spans the 12.03 to 18.19% region of the bacteriophage T7 genome were cloned into pRT29 DNA which is a 5800 base pair plasmid vector containing a single Hpa I site joining pVH51 with a tetracycline resistance element from the TnlO transposon. Detailed restriction mapping of the recombinant plasmids and the pRT29 vector with six different enzymes allowed the identification of two insertions. One recombinant plasmid, pRW307, contained the region between 12.05 and 16.0% of the T7 genome while pRW308 contained the region between 16.1 and 18.19%. pRW307, pRW308, and other recombinant plasmids which were larger than the vector were screened for the origin of T7 replication with an in vitro system specific for the initiation of T7 DNA synthesis. pRW307, but not pRW308 or any other recombinant plasmid, stimulated DNA synthesis approximately ‘I-fold higher than the vector control. This extent of stimulation was comparable to that observed with T7 DNA. Thus, the origin of T7 DNA replication is located between 12.05 and 16.0% of the T7 genome rather than between 16.1 and 18.19%. These results are discussed in relation to the prior electron microscopic studies which located the origin at 17% and the existence of T7 mutants carrying deletions of portions of the 14.6 to 21.8% region.


Replication
of bacteriophage T7 DNA requires both viral and host-specified proteins. The products of viral genes 1 through 6, the products of Escherichia coli genes TsnB and TsnC, as well as DNA gyrase have been implicated in T7 DNA synthesis (l-4). Electron microscopy of replication intermediates from T7-infected E. coli has revealed that T7 DNA replicates as a linear molecule, at least through the first round of synthesis (5). DNA synthesis begins at a single origin located at approximately 17% of the genome and proceeds bidirectionally (6). However, these electron microscopic results are in apparent conflict with the fact that T7 mutants exist which contain deletions in the 14.6 to 21.8%' region (6).
In uitro, T7 DNA synthesis can be studied with a cell-free system prepared from T7-infected E. coli (7,8). This system specifically utilizes T7 DNA as a template and synthesizes intact, infectious T7 DNA molecules (9). Bacteriophage T3 DNA and, to some extent, X-DNA can also function as a template in this in vitro system, but a number of other DNAs are inactive (7,8). The Hpa I restriction enzyme cleaves T7 DNA into 19 fragments that have been mapped (10). One of these, Hpa I Fragment G, extends from 12.03 to 18.19% of the T7 genome (10) and, thus, should span the region where the origin of DNA replication is located. Hpa I Fragment G was partially purified and was cloned using pRT29 (11) as a vector. pRT29 is comprised of pVH51 DNA and a tetracycline resistance element from the transposon TnlO (11). This DNA was employed, for the first time, as a cloning vector since it (a) contains a single Hpa I site (as well as a single Kpn I site) and (b) confers tetracycline resistance to the host cell. A detailed restriction map of this vector is described here. The clones were screened on the basis of plasmid size and two plasmids, pRW307 and pRW308, were found by restriction analysis to contain those parts of Hpa I Fragment G which span the region between 12.05 and 16.0% and 16.1 and 18.19% of the T7 genome, respectively. Using the in vitro system from T7-infected E. coli, it was found that only pRW307 DNA (but not pRW308, the vector, or any other composite plasmid) was capable of strongly stimulating T7 DNA synthesis. These results indicate that (a) an in vitro functional origin of T7 DNA replication is indeed contained within Hpa I Fragment G between 12.05 and 16.0% of the T7 genome; (b) the presence of this origin is sufficient to determine the specificity of the in vitro system; and (c) the origin can function as an integral part of a foreign circular molecule.
A preliminary report of these results has appeared elsewhere (12 and twice banding the phage in CsCl equilibrium gradients, the DNA was extracted as described (18). pRT29 (11) and all other plasmids were isolated as described (13)

Fragment
G is approximately 2460 bp' in length and spans the region between 12.03 and 18.19% of the T7 genome (10). Fragment G co-migrates on the gel with fragment F and the doublet band is barely distinguishable from Hpa I fragments E and H (10).
To prepare sufficient amounts of Hpa I Fragment G for cloning, 100 pg of T7 DNA was digested with Hpa I and fractionated on preparative cylindrical, 1.2% agarose gels. Fragment G was recovered from the gels as described under "Materials and Methods." The purity of the recovered fragments was determined by analyzing small aliquots on 1.2% agarose gels. Fig. lb shows that the doublet of Hpa I Fragments F and G was isolated practically free of other fragments except for small amounts of Fragments E and H. This mixture was used for ligation to the vector.
The mixture of fragments shown in Fig to the dimer, trimer, and higher order products of the insert, and bands of other high molecular weight products.
Restriction enzyme analysis of recombinant plasmids (described below) showed that the Hpa I sites at the two junctions of the 2460-bp insert and the vector were lost. Hence, studies were performed to determine whether this loss was incurred during the Hpa I cleavage of T7 DNA and/or the ligation of fragment G to the pRT29 vector or whether these deletions were caused by cellular functions after the transformation step. An aliquot of a ligation reaction mixture ( Fig. lc) was redigested with Hpa I and analyzed on an agarose gel. All higher order products were converted back to the monomer (data not shown), indicating that virtually all of the Hpa I sites were intact in the ligated molecules. Thus, the sites were apparently removed in Go.
The ligase reaction products were used for the transformation of E. coli C600 SF& the cells were selected for tetracycline resistance, and colonies were screened for insertions as described under "Materials and Methods." Approximately 4% of the tetracycline-resistant colonies contained plasmids larger than the vector (insertion candidates), whereas 15% contained plasmids smaller than the vector. One of the latter group of plasmids was analyzed by restriction mapping and found to be a deletion product of the vector, as expected on the basis of its size.

Identification of a Plasmid
Carrying the Origin of Replication-seven colonies, which contained plasmids approximately 5800 to 8900 bp in length, were each grown in 400 ml of culture. The plasmid DNAs were purified as described under "Materials and Methods." These DNAs were then assayed for their capacity to stimulate deoxynucleotide incorporation into DNA in the T7-specific cell-free system prepared from T7-infected E. coli as described under "Materials and Methods." ' The abbreviation used is: bp, base pairs.  Table I shows that pRT29 DNA, the 5800-bp vector control, showed a low level of stimulation as observed previously (7,8) for DNA not homologous to T7. T7 DNA gave a 4-fold stimulation over that observed for the pRT29 vector. Of the seven recombinant DNA plasmids tested under identical conditions, all but one (pRW307) had a level of activity comparable to or lower than the vector (Table I). pRW307 DNA stimulated dCMP incorporation by 5-to 6-fold relative to the vector. This level of stimulation was comparable to that found with T7 DNA. 3200 These results indicate that a site capable of functioning as an origin of T7 DNA replication was cloned into pRW307. It was recently reported (23) that T7 DNA fragments generated by random degradation or Hpa I restriction were cloned in pMB9. Some of the cloned segments were capable of supplying T7 functions to infecting mutant phages.
Restriction Analysis of pRT29 Vector-pRT29 consists of a 1950-bp Hpa I restriction fragment of the transposon TnlO inserted into the Hind11 site of pVH51 (3850 bp in length) by blunt end ligation (11). The construction and partial characterization of pRT29 was described (11). However, it was necessary to establish a restriction map of this DNA in order to rigorously characterize the insertion candidates. Since pRT29 was not used previously as a cloning vector, its detailed map was unknown. pVH51 and pRT29 were digested with one or two restriction enzymes and the products were fractionated on polyacrylamide or agarose gels, as described under "Materials and Methods." The known lengths (13) of the fragments produced by restriction of pVH51 with Hue III were used for calibrating polyacrylamide gels, whereas the lengths of the fragments produced by restriction of X plac 5 DNA by Eco RI were used for calibrating agarose gels (21). An example of a typical gel pattern is shown in Fig. 2.
The map of pVH51 for the restriction enzymes Hpa I, Eco RI, HindII, Hue II, Hue III, and Hpa II has been determined (13,24,25). The restriction map of the transposon segment of pRT29 was determined for the same enzymes as follows: Eco RI cleaves pRT29 at two sites, one of which is located inside pVH51 and its exact position is known (24). Consequently, the position of the other site can be determined from the size of Fragment B (Fig. 3 and Table II RI site. Fragment C is not cleaved by Eco RI and therefore must be located between the Eco RI sites. Hpa I cleaves pRT29 only once (11). Hpa I recognizes only one of the four possible sequences recognized by HindII. Thus, cloning of a Hpa I fragment into a Hind11 site regenerates the Hind11 sites, but these sites are not necessarily recognized by Hpa I. In the construction of pRT29, only one Hpa I site was regenerated (11). Digestion of pRT29 with Hpa I and Eco RI shortened Eco RI fragment A but did not cleave Eco RI fragment B indicating that, of all three Hind11 sites present in pRT29, only the site spanned by Eco RI fragment A is also a Hpa I site (Fig. 3).  Fig. 3. Finally, Fragment D is cleaved by Hind11 but not by Eco RI. Some of these data are shown in Fig. 2 (Fig. 2).
The lengths of the fragments produced by restriction of pRT29 by each of these enzymes are listed in Table II. Restriction Analysis of Insertion Candidates-Insertion candidates, recombinant plasmids expected to contain DNA insertions on the basis of their total length, were restriction mapped using the methodology applied to pRT29. pRW307 was found to have a total length of 6250 bp suggesting the presence of a 450-bp insertion. However, restriction with Hue III revealed that Fragments E, N, A, and M of the pRT29 vector were absent from pRW307 and were replaced by two new Hae III fragments which were 1500 and 650 bp in length ( Fig. 3 and Table III). The relative positions of these two fragments were established from double digests of pRW307 with Hue III + Eco RI. pRW307 was not cleaved by Hpa I apparently because the two Hpa I sites, which should have been regenerated at the junction between the T7 Hpa I Fragment G and the Hpa I linearized vector, were deleted. Eco RI and Hue II cleave pRW307 only at the sites present in the vector portion of the molecule.
Hind11 cleaves pRW307 at three sites inside the insertion generating fragments A', B', C', and D' (Fig. 3). The positions of Fragments A' and B' were determined from double digests with Hind11 + Hue III. The positions of Fragments C' and D', relative to one another, were determined from partial digests of pRW307 with HindII. The only orientation compatible with all partial restriction products is shown in Fig. 3. In order to determine what length of Hae III Fragment E of the vector had been replaced by the insertion, Hue III Fragments A' and B' of pRW307 were purified by RPC-5 column chromatography; the elution profile and gel analysis are described elsewhere (26). Fragment A' was then digested with Ah I. The restriction map of Fragment E with AZu I is known (13). Digestion of Hae III fragment A' with Ah I revealed that all of Band E is present in pRW307 except for a 4%bp fragment proximal to Fragment N. Therefore, the insertion must end within this fragment.
The results of the restriction analysis of pRW307 described above led to the construction of the map shown in Fig. 3. In pRW307, approximately 1200 bp were deleted from the vector (Hue III Fragments A and N, and portions of M and E), whereas a T7 DNA fragment approximately 1600 bp in length was inserted.
In order to establish that the DNA insertion in pRW307 in fact spans part of Hpa I Fragment G, the purified Hue III fragments A' and B' of pRW307 were digested with Hpa II. It is known that two of the restriction fragments generated by digestion of T7 Hpa I Fragment G with Hpa II are 140 and 138 bp in length and the 140-bp fragment is located immediately to the left of the 13%bp fragment.' The latter is cleaved by Hae III into two fragments, 115 and 23 bp in length (27)   tions of Hpa II Fragments B' and C', as well as the position of the 140-bp fragment to the left of the 138bp fragment, were verified with a double digest of Hue III fragment A' with Hpa II + HindII.
In this way, the single Hue III site inside the insertion of pRW307 was positively identified as the Hue III site at approximately 14.6% of T7 DNA. In addition, the position of Fragments 140 and 138 relative to one another allows orientation of the direction of the insertion relative to that of T7 DNA. The direction shown on the map of Fig. 3 corresponds to the left-to-right direction of T7 DNA. Therefore, by considering the lengths of Hue III Fragments A' and B', it was determined that the insertion is spanning the region between 12.05 and 16.0% of T7 DNA.
Further evidence that the DNA insertion in pRW307 is indeed the 12.05 to 16.0% region of T7 DNA was provided by a comparison of the map of Fig. 3 with the restriction map of the left quarter of T7 DNA that was determined by Studier.' The two maps were in complete agreement on the positions of all of the fragments shown in Fig. 3, leaving little doubt about the identity of the insertion in pRW307. In addition, the sequence of portions of Hpa II Fragment E' have been determined" and are in complete agreement with the published sequence (27).
The insertion candidate pRW308 was restriction-mapped in ' N. Panayotatos and R. D. Wells, unpublished results. the same manner as described above. It is approximately 6800 bp in length, contains an insertion of 840 bp, and has retained one Hpa I site. The insertion in pRW308 contains no other Hind11 site other than the one recognized by Hpa I. Restriction with either Hue III or Hpa II produces three fragments of the sizes shown in Table IV. In addition, it was determined that only a single Hha I site is present inside the insertion in pRW308 at approximately 150 bp to the left of the Hpa I site. The restriction map for the insertion in pRW308 is shown in Fig. 4. Comparison of this map with the map determined for the left quarter of T7 DNA by Studier' reveals that the insertion in pRW308 must originate from the region between 16.1 and 18.19% of the genome. Thus, pRW308 contains an insertion which spans that part of Hpa I Fragment G that was deleted in pRW307, except for approximately 40 bp at 16.0 to 16.1% which do not appear to be present in either insertion.
Finally, five more plasmids (pRW303, pRW304, pRW309, pRW310, and pRW311) were judged to contain DNA insertions. pRW304 is approximately 6600 bp in length and carries an insertion of 900 bp but no Hpa I sites. The other plasmids varied in size from 6650 to 8900 bp, but their restriction patterns were too complicated for a restriction map to be constructed.
We assume that these are also insertions of T7 DNA.
Further Characterization of the Template Properties of the Plasmids--In order to further characterize the degree of specificity of the in vitro system for T7 DNA and pRW307, the dependence of DNA synthesis on time and DNA concentration was compared with that of the pRT29 vector. The results (Fig. 5, A and B) show that pRW307 and T7 DNA exhibit a similar dependence on time and DNA concentration and both are more efficient templates than pRT29. DNA synthesis increases linearly for approximately 10 min and then gradually declines. At 30 pg/ml, DNA synthesis supported by pRW307 and T7 DNA is stimulated 7-and 8-fold over that of the vector, respectively.
The rate of incorporation for a 20min reaction is approaching saturation for a template concentration of approximately 40 pg/ml.  The restriction fragments of the vector that flank the insertion are shown by dashed lines, designated by unprimed capital letters, and correspond to the vector fragments in Fig. 3 The observed dependence on time and template concentration for T7 DNA is in good agreement with the published data (7,8). Thus, the fact that pRW307 behaves similarly to T7 DNA is consistent with the notion that the specificity and efficiency of the in vitro system depends on the presence of a functional origin of T7 DNA replication.
The results shown in Fig. 5B indicate that the total amount of DNA synthesized represents 10 to 20% of the input T7 template. Under the same conditions, a 2-fold higher productto-template ratio was obtained with a cell-free system prepared from E. coli DRllO (a pal A' host) infected with bacteriophage T7s,sm (a mutant lacking the exonuclease product of gene 6) (9). Masker and Richardson have shown that the T7 DNA product synthesized with this system is not covalently attached to the template and up to 50% can be of the size of intact T7 DNA (9). In a preliminary characterization of the nature and length of the DNA synthesized with pRW307 as the template, the products were labeled with (Y-["2P]dCTP as substrate, restricted with Hoe III and HindII, and fractionated on 5% polyacrylamide gels. The bands of unlabeled DNA were stained with ethidium bromide while the radioactive product was visualized by autoradiography. Identical restriction patterns were observed with the two methods indicating that the 32P label becomes incorporated into double-stranded molecules that can be restricted to completion." After longer incubation periods (1 min), all restriction fragments appeared to be uniformly labeled. However, on short incubation times (10 s), a higher proportion of the label was associated with the restriction fragments that span the T7 insertion, as evidenced from the relative intensity of the bands on the autoradiogram.
The cell-free systems used in these studies are also highly specific for the initiation of T7 RNA synthesis catalyzed by the co-purified T7 RNA polymerase (7,8). In the presence of rifampicin, which inhibits bacterial RNA polymerase but not the T7 enzyme, incorporation of radioactive label from a ribonucleoside triphosphate into an acid-insoluble product proved to be a powerful screening procedure for the presence of T7 late promoters in the recombinant plasmids. In fact, pRW307 stimulated RNA synthesis approximately 300-fold higher than pRW308, the vector, or any other recombinant plasmid." This result indicates that a late promoter(s) is present in the 12.05 to 16.0% region (in agreement with the recent identification of such a promoter at approximately 14.5% (27), but not in the 16.1 to 18.19% region. The identification and characterization of the promoters cloned in pRW307 are currently under investigation. DISCUSSION Two segments of the 2460-bp Hpa I Fragment G (12.03 to 18.19%) were cloned and used to localize and further charac-terize the T7 origin of replication.
One plasmid (pRW307) contained a 1600-bp insert corresponding to 12.05 to 16.0% and another (pRW308) contained an 840-bp insert corresponding to the T7 region between 16.1 and 18.19%. The DNA template properties of these and other plasmids were determined using the highly specific and efficient cell-free system (7,8) from T7-infected cells which is capable of synthesizing infectious T7 DNA de nova (28). Only the plasmid (pRW307) containing the T7 segment from 12.05 to 16.0% specifically stimulated DNA synthesis. Hence, we conclude that the primary origin of replication is located between 12.05 and 16.0% of the T7 genome.
Prior electron microscopic studies (6) on T7 replication intermediates formed in uiuo located the origin at 17% of the genome. Further similar studies4 refined the origin position to 16.5% with an estimated error of +l.O to 1.5% of the genome. Hence, it is likely that the refined position agrees with our studies. Likewise, both the electron microscopic work and our cloning investigations may be consistent with the existence of T7 mutants for which portions of the region between 14.6 and 21.8% and thus the origin (?) had been deleted' (29). To explain this contradiction, Dressler et al. (6) previously proposed that these mutants may use a secondary origin. However, if the T7 origin is at approximately 14.5%, this region would be present in the deletion mutants, would be just within the error of the electron microscopic measurements, and would be contained by pRW307. Preliminary results" with recombinant plasmids containing portions of the 12.05 to 16.0% region are in agreement with the conclusion that the origin is at -14.5%.
Casual consideration of the data presented in Fig. 4 would seem to support the notion that the segment cloned in pRW307 can initiate DNA synthesis almost as efficiently as the origin in T7 DNA. However, a direct comparison is not possible at present since the DNA synthesis studies (Table I and Fig. 5) measure deoxyribonucleotide incorporation and not the frequency of initiation events. pRW307 is approximately 6 times smaller in size than T7 DNA. Because of this difference, the concentration of pRW307 "origin" is approximately 6 times higher than the T7 "origin," at equal nucleotide concentrations.
On the other hand, bidirectional replication can, at least in theory, proceed 6 times farther on T7 DNA than on pRW307 before termination.
Thus, although a higher "origin" concentration is present for pRW307, its activity may be severely limited by early termination.
Therefore, in reactions which measure total dNMP incorporation, a direct comparison of the rates of DNA synthesis supported by templates (such as pRW307 and T7 DNA) that have large size differences is not very useful. However, such comparisons are valuable for templates such as pRW307, the vector, and the other recombinant plasmids shown in Table I which are similar in size.
From the studies described herein, it is clear that the T7 origin is functional in vitro as part of a circular recombinant molecule. The circular pRW307 DNA is not linearized by a contaminating activity in the in vitro replication system" even though a powerful topoisomerase is present. Other studies (5) have indicated that T7 DNA replicates in viva as a linear molecule without going through a circular intermediate for, at least, the fist round of synthesis. A DNA gyrase, however, appears to be necessary for T7 DNA replication in viuo (4). Restriction analysis of pRW307 indicated that a considerable amount of DNA was deleted from both the vector part and T7 DNA. Approximately 1200 bp were deleted from the vector but almost none from the inserted fragment at their approximately 1000 bp of T7 DNA but almost none of the vector were deleted at their junction inside Hue III Fragment B'. One might argue that the DNA was lost through random degradation before the ligation reaction occurred. However, the "sidedness" in the deleted DNA as well as the retention of the Hpa I termini during the in vitro ligation reaction (Fig. lc) suggest that these deletions occurred during or after the cell-transformation step. The possibility exists that the inserted T7 fragments might have suffered additional internal deletions. However, the complete agreement between our data on the positions of 14 restriction sites distributed throughout the insertions in pRW307 and pRW308 and the data of Studier' for these sites in T7 make this possibility unlikely.
The possibility exists that the high level of DNA synthesis observed with pRW307 relative to pRW308 and the pRT29 vector was not due to the T7 DNA insertion, but rather due to the deletion of a hypothetical replication control element located on the part of pRT29 that was deleted in pRW307. Evidence against this possibility comes from the following experiment: Hind11 Fragment A' (Fig. 3) was isolated and recircularized." The resultant plasmid (pRW312) contained all of pVH51 (except for the portion deleted in pRW307) and approximately 50 bp of the T7 DNA insertion (Fig. 3). pRW312 was assayed with the cell-free system and was found to have activity comparable to pVH51 and pRT29. The lack of stimulation of DNA synthesis by pRW312, which carries the same deletion with respect to its parental vector as pRW307, indicates that the stimulation observed with the latter plasmid is not due to the deletion of an inhibitory element, but rather due to the T7 insertion.
It was recently reported that Hpa I Fragment G of T7 DNA was cloned in pMB9 along with a number of other T7 DNA fragments generated by Hpa I restriction or by random degradation (23). Hpa I Fragment G was not present intact in that clone either but contained an insertion of some DNA segment at approximately 16.0%.' It may be fortuitive that the insertion in pRW307 ended at 16.0%, whereas the insert in pRW308 began at 16.1% of the T7 genome. However, it is possible that a recombination "hotspot" is located at 16.0 to 16.1%. DNA sequencing studies on this region may be revealing.